Forecasting on Kalshi August 2025 resolution filter (OOD)
0.228Brier ScoreBBC
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| BBCInput LLM=Qwen3-32B2026.05 | 0.228 | 60.9 | 0.658 | 0.059 | |
| BBCInput LLM=Qwen3-8B2026.05 | 0.235 | 59.9 | 0.62 | 0.061 | |
| VerbalizedInput LLM=Qwen3-32B2026.05 | 0.238 | 60.5 | 0.651 | 0.097 | |
| Isotonic RegressionInput LLM=Qwen3-32B2026.05 | 0.244 | 60.5 | 0.651 | 0.141 | |
| OpenForecaster-8BInput LLM=Qwen3-8B2026.05 | 0.244 | 59.9 | 0.632 | 0.093 | |
| Platt ScalingInput LLM=Qwen3-32B2026.05 | 0.251 | 59.6 | 0.651 | 0.148 | |
| Future-as-a-label-32BInput LLM=Qwen3-32B2026.05 | 0.258 | 60.7 | 0.638 | 0.159 | |
| VerbalizedInput LLM=Qwen3-8B2026.05 | 0.258 | 58.5 | 0.609 | 0.116 | |
| Platt ScalingInput LLM=Qwen3-8B2026.05 | 0.258 | 59.5 | 0.609 | 0.146 | |
| Isotonic RegressionInput LLM=Qwen3-8B2026.05 | 0.258 | 59.7 | 0.608 | 0.152 |